Fleet K. The Cambridge History of Turkey: Volume 1, Byzantium to Turkey, 1071-1453

Подождите немного. Документ загружается.

machiel kiel

a number of Islamic buildings and the place where he died and was buried.

Some of his works in Yenice still survive today.

The years between 1371 and

1383 are thus the most likely for the G

ulcine buildings, between 1383 and

1390 for that in Serres and the last decade of the fourteenth century for the

works in Yenice-i Vardar.

The Evrenos

Imareti in G

ulcine is a good representative of the zaviye-

T-plan mosque which can better be called a hanekah (dervish lodge), or zaviye

(dervish lodge). It does not face towards Mecca, and is composed of a sizeable

rectangular hall, covered by a dome on a network of Turkish triangles and

by a barrel vault over the remaining part. The hall opens up towards the east

with a huge arch of soft yellow-green limestone, emphasising the entrance,

the remainder of the work being made up of a cloisonn

e of bricks and broken

stone. The central hall is ﬂanked on both sides by transversely placed rectan-

gular side rooms which are accessible only from the central hall by means of a

door, framed by ﬁne portals. The one on the left (south) side has a large Arabic

inscription, calligraphed and cut into the arched tympanon above the actual

entrance, the whole being a part of the authentic structure and not a work

of the recent restoration. The inscription is evidently the ‘Bauinschrift’ of Gazi

Evrenos’s time but it is illegible, having been mutilated by the Bulgarian con-

querors of 1912, who identiﬁed the building as a genuine old-Bulgarian church,

‘reconverted’ it into a church and dedicated it to the Holy Tsar Boris-Michael.

The letters of Gazi Evrenos’s inscription were cut away to be replaced with a

Cyrillic inscription. The modern Greek restorers carefully cleaned the remains

of the original inscription which remains, alas, for the greater part illegible.

It is the oldest monument of Ottoman epigraphy in south-eastern Europe.

Against previously held theories, the building never had a domed or arched

porch. When, during the restoration, the ugly concrete machinery halls were

removed (before the restoration the building had served as the town’s power

station) and archaeological research was carried out, no trace of a portico of

any type was found. The form the original building had, and now has again,

ranks it with the early fourteenth-century, pre-Ottoman zaviyes like those in

Tokat, rather than associating it with the earliest Ottoman zaviye-mosques

such as Orhan Gazi’s in Bursa (1335). The G

ulcine plan is also partly related

to the Nil

ufer

Imareti in

Iznik (1388), especially the rectangular, partly domed,

partly arched lateral guest rooms. Like G

ulcine, the

Iznik building does

59 See M. Kiel, ‘Yenice-i Vardar (Vardar Yenicesi – Giannitsa): a Forgotten Turkish Cultural

Centre in Macedonia of the 15th and 16th Century’, Studia Byzantina et Neohellenica

Neerlandica 3 (1971), 300–55; more easily available in M. Kiel, Studies on the Ottoman

Architecture of the Balkans (Aldershot, 1990).

160

Ottoman expansion into the Balkans

not face towards Mecca. The link between the early zaviyes in Tokat and the

ulcine building is indirect. The direct link is missing, but must have been

formed by buildings which have now disappeared.

A hint of the real nature

of this and other buildings (Serres) within this group is given in the survey of

staff, income and expenditure of the G

ulcine building contained in the reg-

ister from 858/1454–5. The most important and best paid member of the staff

is called an ‘Ahi’, thus a member of the famous late Seljuk and early Ottoman

brotherhood of the ahis.

The Eski Cami of G

ulcine is preserved in a much

altered and enlarged form. It stands slightly east of the imaret and consists of

an original fourteenth-century domed cube with an inner space of 7.80 metres

which in the nineteenth century was enlarged by a spacious addition, envelop-

ing it on most of its three sides. The original entrance wall was then removed

to join both old and new sections; a building inscription over the door was lost

in the process. The building was evidently designed as a mescid (small mosque)

in the manner of many other single-domed prayer chapels, as preserved in

their dozens in old towns such as Afyon or Ayasolu

g/Selc¸uk and elsewhere

in Anatolia. Both G

ulcine buildings are closely linked by their plans with

contemporaneous structures in Anatolia, from where the inspiration came.

They are built, however, using the late Byzantine cloisonn

e technique, then

in vogue in the southern Balkans as well as in north-western and Aegean

Anatolia, where we can assume the participation of local masters.

The old-

est preserved Ottoman utilitarian building in the Balkans also goes back to

Gazi Evrenos. It is the han near the mineral springs of Ilıca (Trajanopolis),

60 For the Tokat zaviyes see the excellent work of Sedat Emir, Erken Osmanlı Mimarlı

gında

C¸ ok-

Is¸levli Yapılar: Kentsel Kolonizasyon Yapıları Olarak Zaviyeler, vol. II:

Onc

ul Yapıları

Tokat Zaviyeleri (Izmir, 1994). For the zaviye-mosque see also the inﬂuential article of

Semavi Eyice, ‘

Ilk Osmanlı Devrinin Dini-

Ic¸timai Bir M

uessesi: Zaviyeler ve Zaviyeli

Camiler’,

Istanbul

Universitesi

Iktisat Fak

ultesi Mecmuası 22, 1–2 (1962–3), 1–80.

Istanbul Belediye K

uphanesi, Cevdet Yazmaları collection, no. o.89. For this broth-

erhood and its role see ‘Akhi’, in The Encyclopaedia of Islam, 2nd edn (Leiden, 1960–

2006) [henceforth EI2], i,pp.321–3 (Franz Taeschner), or Ziya Kazıcı, ‘Ah

ılik’, in TDV

IA,

vol. I (Istanbul, 1988), pp. 540–2. For more details see Nes¸et C¸a

gatay, Bir T

urk

Kurumu

olan Ah

ılik (Ankara, 1974).

62 For both buildingsof Gazi Evrenos in Komotini seeM. Kiel, ‘Observationson theHistory

of Northern Greece during the Turkish Rule: Historical and Architectural Description

of the Turkish Monuments of Komotini and Serres, their Place in the Development of

Ottoman Turkish Architecture and their Present Condition’, BalkanStudies 12 (1971), 415–

44; and Kiel, ‘The Oldest Monuments of Ottoman-Turkish Architecture in the Balkans:

the

Imaret of Gazi Evrenos in G

ulcine (Komotini) and the Evrenos Bey Khan in

the Village of Ilica/Loutro in Greek Thrace (1370–1390)’, Sanat Tarihi Yillı

gı 12 (1983),

117–38. See also the report on the inspired restoration of the building by the men who

undertook it: Charalambos Bakirtsis and Pantelis Xidas, ‘Imaret. Komotini, Greece’, in

Secular Medieval Architecture in the Balkans, 1300–1500, and its Preservation, ed. Slobodan

Cur

c and Evangelia Hadjitriphonos (Thessalonike, 1997), pp. 294–5.

161

machiel kiel

15 kilometres east of the modern Greek port of Alexandropolis, on the road to

Ipsala on the Turkish side of the Maritsa (Evros) river. The han is a spacious,

barrel-vaulted, rectangular hall with ﬁreplaces along the southern wall. This

hall is preceded by a closed entrance room, now largely ruined. As a type it is

closely related to a whole group of hans from the Beylik period in Anatolia.

The institution of han or caravansary was unknown to Byzantino-Slavic archi-

tecture. An inscription, mentioning the name of Gazi Evrenos as its founder,

existed, until the collapse of the fac¸ade in 1940, above the entrance and was

mentioned by the local historian A. Samothrakis.

The early seventeenth-

century historian of Edirne, Hibri Efendi, described it and also attributes it to

Gazi Evrenos. It must have been constructed some decades after Evrenos’s

buildings in G

ulcine, thus in the 1390s.

The oldest representative of commemorative architecture is apparently the

urbe, mausoleum, attributed to Lala S¸ahin Pas¸a in Kazanl

ak. It lies on a spur

of the Balkan mountains, overlooking the town of Kazanl

ak in Bulgarian

Thrace. This town ﬁgures as Akc¸e Kazanlık in the early Ottoman chronicles of

the conquest of the Balkans, and appears as a small, purely Turkish settlement

in the oldest preserved Ottoman registers for this part of Thrace. The building

is an early example of the open canopy t

urbe with a dome resting on four

brick-built pointed arches and four square pillars, likewise of brick. Lala S¸ahin,

the ﬁrst Ottoman beylerbey of Rumeli, died shortly after 1383. His body was

buried in his still-standing t

urbe in Kirmasti (now Mustafa Kemal Pas¸a) 70

kilometres west of Bursa.

His bowels must have been buried in the hill

above Kazanl

ak, most probably the place where he died. The name of the

building was handed down by tradition among the Muslim inhabitants and

recorded by Franz Babinger in the 1930s. An interesting early parallel to the

habit of burying the intestines of an important person on the spot where he

died and preparing (or mummifying) his body for the long transport to his

desired burial place is handed down in the chronicle of the so-called Serbian

janissary, Konstantin of Ostrovi

c, who had entered Ottoman service in 1455,

after Mehmed II’s capture of the Serbian silver-mine town of Novobrdo on the

63 A. Samothrakis, in Thrakika Chronika 18 (1943). A large fragment of this inscription was

found recently and is now kept in the Archaeological Museum of Kavalla. In the text it

is mentioned that Evrenos Bey was a hacı. It is known from early Ottoman chronicles

that Evrenos Bey asked Bayezid I for permission to go on hac in 1396. The building was

thus constructed in the last years of the fourteenth century. The important text of the

inscription is to be published by Heath Lowry.

64 For this building see Kiel, ‘Observations’ and Kiel, ‘The Oldest Monuments’; and Heath

W. L o w r y , The Shaping of the Ottoman Balkans 1350–1550 (Istanbul, 2008), pp. 29–35.

65 See Ekrem Hakkı Ayverdi, Osmanlı Mim

ısinin

Ilk Devri 630–805 (1230–1402), vol. I

(Istanbul, 1966), pp. 190–7, with good plans and photographs.

162

Ottoman expansion into the Balkans

edge of the plain of Kosovo. This source notes that, after the battle of Kosovo

in 1389, ‘Sultan Bayezid, after having been victorious, remained on the plain

of Kosovo and erected there, on the battleﬁeld, over the site where his father

was killed, a monument: on four columns a dome was made and was covered

with lead, and this stands till the present day.’

The Serbian janissary probably

passed this monument when on campaign before 1463, when he was taken

prisoner by King Mattias Corvinus, before ﬁnally settling in Poland, or knew

it from his youth as he was born in the close vicinity of the monument. The

same story, in different wording, is found in the ﬁfteenth-century Ottoman

chronicle the Anonymous-Giese where, after the description of the battle of

Kosovo, the author states: ‘Then they built on the side where he fell a t

urbe

and made the place visible. Even now this t

urbe remains on this site. Then they

transported his body to Kaplıca [suburb of Bursa, where Murad’s principal

buildings are located] and buried it there.’

The t

urbe was later replaced by

a more monumental structure. The one we still see today was made on the

order of Sultan Mehmed Res¸ad, a year before the end of Ottoman rule in the

Balkans (1911). When Sultan S

uleyman died before Szigetvar in 1566, the same

procedure was followed, the t

urbe later being surrounded by a garden and a

zaviye added to it. S

uleyman’s body was buried in Istanbul, as is known; that

of Murad I in Bursa in a monumental t

urbe. The practice is deﬁnitely older,

and must be connected to the pronounced ancestor cult of the Turkic people,

in the case of Kazanl

ak and Kosovo, visibly transported to the Balkans.

Among the very early dated Ottoman single-domed mosques ranks the Eski

Cami in Kırkkilise in Turkish Thrace. It is a cube covered by a dome over an

66 Konstantin of Ostrovi

c, Memoiren eines Janitscharen: oder T

urkische Chronik, trans. Renate

Lachmann (Graz, Vienna and Cologne, 1975), p. 82.

67 Giese, Altosmanischen Anonymen Chroniken,p.38.

68 On the practice of mummifying see Faruk S

umer, ‘The Seljuk Turbehs and the Tradition

of Embalming’, in Atti del secondo congresso internationale di arte turca (Naples, 1965),

pp. 245–8. For the wider background, V. Barthold and M. J. Rogers, ‘Burial Rites of

Turks and Mongols’, Central Asian Journal 19 (1970), 195–227.OnKazanl

ak and Lala

S¸ahin see Franz Babinger, Beitr

age zur Fr

uhgeschichte der T

urkenherrschaft in Rumelien

(14.–15. Jahrhundert) (Brno, Munich and Vienna, 1944), pp. 67, 72–4, which includes

a rich bibliography; Franz Babinger, ‘Beitr

age zur Geschichte des Geschlechtes der

Malko

c-oglu’s’, in his Aufs

atze und Anhandlungen zur Geschichte S

udosteuropas und der

Levante, vol. I (Munich, 1962), pp. 355 ff. and illustr. 44, giving the pre-war situation of the

Kazanl

ak monument. For a recent photograph see M. Kiel, ‘Osmanische Baudenkm

aler

inS

udosteuropa,Typology undVerh

altnisszur Lokalen Kunst– ProblemederErhaltung

in den heutigen Nationalstaaten’, in Die Staaten S

udosteuropas und die Osmanen, ed. Hans

Geor

Majer (Munich, 1989), p. 59.InJuly1998 the condition of the monument had

largely remained the same. The t

urbe of Murad I on Kosovo Polje was recently superbly

restored with help from Turkey. See the ﬁne publication of Mehmed

Ibrahimgil and

Neval Konuk, Kosova Sultan Murad H

udavendig

ar T

urbesi (Ankara, 2005).

163

machiel kiel

inner space of 12.25 metres, the dome resting on a belt of Turkish triangles. The

original inscription is lost but one recording a major repair mentions Hızır Bey,

an earlymember of the well-known Mihalo

glu family, as founder,and 785/1383–

4 as the date of construction.

The early Ottomans quickly developed the

small, cubic single-domed mescids they had inherited from the Seljuks into

Friday mosques (the Cami mosques) of much more monumental proportions

and artistic pretensions. It was this type which became very popular in the

Balkans, in most places the embodiment of all that was regarded as ‘Ottoman’

and in most cases the sole representative of this style. Gazi Evrenos’s single-

domed building in G

ulcine was doubtlessly conceived as a mescid.The

Eski Cami of Serres, originally also a single-domed building, was deﬁnitely

a mosque from the beginning, as stated in its building inscription. It was

the largest mosque of the city for more than a century, until 1492, when the

huge mosque of Mehmed Bey, the son of the famous Gedik Ahmed Pas¸a, was

built.

The Eski Cami of Serres was demolished in 1938 but old photographs and the

topographical plan of the city from the inter-war period, as well as a description

by Evliya C¸ elebi, give enough data to reconstruct it. It was a foundation of

the conqueror of the city, grand vezir C¸ andarlı Hayreddin Pas¸a, who erected

it between 1383, the year in which the city was taken, and 1387, the year of his

death, just outside the city walls of Byzantine Serres. As a building it appears

to be closely related to Hayreddin’s more famous creation, the Yes¸il Cami in

Iznik, the seat of the C¸ andarlı family.

Evliya C¸ elebi gives the date of its completion as 787/1385 and describes it

as the ﬁrst mosque of the city. It was a rectangular prayer hall covered by one

large dome of nearly ten metres span, and, as in

Iznik, by three vaults resting on

two pillars. It was preceded by a domed porch. This solution reminds us of the

narthex and exo-narthex of the middle Byzantine churches and might indeed

be a creative reinterpretation of it. In the eighteenth century the building was

enlarged on both lateral sides, and a second porch was added. Soon after its

initial construction the Eski Cami was to become the focal point of the city.

More than a century after its construction C¸ andarlı

Ibrahim Pas¸a constructed

the six-domed bedestan of Serres just opposite the mosque of his great ancestor,

and thus became the focal point of the city’s economic life. The old Byzantine

town, however, with its dozens of churches, became largely a residential area,

69 For this mosque and its inscription see Mehmet Tuncel, Babaeski, Kirklareli ve Tekirda

Camileri (Ankara, 1974), pp. 23–5;

Ozcan Mert, ‘Kirklareli Kit

abeleri’,

Istanbul

Universitesi

Edebiyat Fak

ultesi Tarih Dergisi 25 (1971), 155–62.

164

Ottoman expansion into the Balkans

where the Christians continued to live. The Serres bedestan has managed to

survive until today, as has been mentioned above.

At ﬁrst sight the important Eski Cami of the town of Yambol on the Tunca

(Tund

za) river in Bulgarian Thrace appears to be an early endeavour to develop

the plan of an Ottoman great mosque of the kind of the

Uc¸S¸erefeli in Edirne.

In fact it is the work of two different building campaigns, as became visible

during its restoration when an ugly coat of nineteenth-century plaster was

taken off. The building is very early. Yambol, the Byzantino-Bulgarian border

town of Diampolis, was conquered by the Ottomans in the summer of 1370.

It capitulated after a long siege. The small and very strongly walled town in a

loop of the Tunca must have been left as it was. It was still defendable when

in 1412 Musa C¸ elebi attacked it. After that date the town walls must have been

left to decay and disappear slowly, as was the case in Serres, and in other places

where they were no longer needed. The Ottomans constructed their early

buildings in the open plain to the north of the old town. The most important

is the Eski Cami, of which the nucleus reaches back to the 1370sor1380s.

This part of the building is a good example of a single-domed mosque with

engaged pillars. These pillars, connected by heavy, pointed arches, reduce the

space which needed to be covered by the dome, but at the same time create

a richer and wider inner space. The direct forerunner of this type of mosque

seems to be the mosque of Koca Yadıgar in the town of

u near Eskis¸ehir in

western Anatolia, said to date from 776/1374. An older, slightly larger and more

convincing example is the mosque of Orhan Bey in Eski Bilecik, an undated

building but certainly constructed before the death of this second ruler of the

house of Osman (1362). Both the Bilecik and

u buildings have no porch.

Yambol, however, has a massive and monumental porch with closed ﬂanks

and an arcade covered by crossvaults resting on three arches carried by two

square pillars. It is a solution which we also encounter in some thirteenth-

century Seljuk mescids in Konya. To the right of the entrance rises a curious

square minaret, the only one known in the Balkans, but rather common in

south-eastern Anatolia, in the art of the various Turkoman beyliks, and clearly

connected with the Syrian prototype. Locally the minaret was long held to be

the tower of the old Bulgarian cathedral of Yambol but archaeological research

70 A number of old photographs of the Serres Eski Cami as it was in the ﬁrst decades of

the twentieth century were published by G. Kaftantsis,

    



, (Thessalonike, 1985). See also Th. Pennas,

  



 ! "   #$ %  &'   

 ()

, 1383–1913, 2nd edn (Athens, 1966). For the functioning of the vakıf of this

mosque see the useful work of Balta, Les vakifs de Serr

es,pp.91–5.

165

machiel kiel

underneath the ﬂoor of the mosque found no trace of such an imagined

building.

Many decades later, owing to the rapid expansion of the Muslim population

of the town (the 1454–5 register mentions 591 adult Muslim males but only 56

adult male Christians)

and perhaps inspired by the great

Uc¸S¸erefeli Cami

in Edirne (1437–48), the Yambol mosque was enlarged on two sides by three

large units on each side, covered by cradle vaults. In this manner a building

was created which evidently belongs to the same group as the great Edirne

mosque.

From all the works erected by Gazi Evrenos in his residence, the new town

of Yenice-i Vardar, only a large single hamam remains. It shows a complicated

plan with a profusion of domes on belts of elaborate Turkish triangles but

does not have the later obligatory tripartite layout. Like some other early

Ottoman hamams, or those from the second half of the fourteenth century in

the west Anatolian princely residences of Ayasolu

g and Balat, it has no large

disrobing room. The Yenice-i Vardar building has been a ruin for decades, but

is protected by law against demolition. It is not dated but has to be placed in

the 1390s. In that decade Evrenos had moved his seat again and created a new

town, probably on the site of a small Byzantine castle (the name Yenice does

not mean new, but newish). His foundations there consisted of a mosque, an

imaret, a great medrese (school) and the preserved hamam. His buildings are

described by Evliya C¸ elebi and are mentioned in the various tahrir deftersof

the sixteenth century. Gazi Evrenos’s t

urbe, near the hamam, survived in the

form of an early twentieth-century reconstruction. The building inscription,

mentioning the death of the old warlord in 1417, was found some decades ago

and was published by Vasili Demetriadis.

From the same time as Gazi Evrenos’s buildings in Yenice-i Vardar, the last

decade of the fourteenth century, are two buildings in the small central Bul-

garian town of

Ihtiman.

Ihtiman was founded by members of the Mihalo

glu

Istanbul Belediye K

uphanesi, Cevdet Yazmaları collection, no. o.89 (ah 859/1454–5).

72 For an older description of the Yambol Eski Cami see M. Kiel, ‘Some Early Ottoman

Monuments in Bulgarian Thrace, Stara Zagora, Jambol and Nova Zagora’, Belleten 38,

152 (1974), 644–9; updated in Kiel, Studies, section VII. A detailed survey of the results

of the great restoration of this building and the archaeological research underneath its

ﬂoor has not been published. For Orhan Bey’s mosque in Eski Bilecik and Koca Yadıgar

u see Aptullah Kuran, The Mosque in Early Ottoman Architecture (Chicago, 1968),

p. 70, with plan and photograph.

73 For the history and the monuments of this town see Kiel, ‘Yenice-i Vardar’, pp. 300–29;

also Kiel, Studies; Vasili Demetriadis, ‘The Tomb of Ghazi Evrenos Bey at Yanitsa and

its Inscription’, Bulletin of the School of Oriental and African Studies 39 (1976), 328–32;and

Lowry, The Shaping of the Ottoman Balkans,pp.58–64.

166

Ottoman expansion into the Balkans

family, three kilometres south of the Byzantino-Bulgarian town of Stipion

(Stiponje), as the centre for a group of Turkish villages in this small highland

valley which is surrounded on all sides by heavily wooded mountains but con-

trols the main road from Edirne to Soﬁa and further on to Belgrade. After the

conquest, Stipion long remained the largest, and purely Christian, settlement

in the region. It was taken by treaty and the name

Ihtiman refers to this (‘ahd

ve aman’). The Mihalo

glu constructed an imaret/zaviye as focal point of the

new settlement and nearby a han of the kind found in Ilıca (Trajanopolis) and

a hamam. The tax revenues of the Turkish villages of the mountain plain, and

that of Stipion, were endowed for the upkeep of the foundation. Its founder

appears to be Mihalo

glu Mahmud Bey, who fell in the battle of Ankara in 1402.

The imaret/zaviye, for half a century in an incredible state of ruin, consists

of a central domed space ﬂanked by lateral rooms, also covered by domes,

and a barrel-vaulted section which houses the mihrab. Originally the lateral

spaces were closed and only opened to the central space through doors. In the

course of the sixteenth century, when the old institution once housed in the

building no longer existed, the separating walls were removed and the space

incorporated to the functions of the mosque. To the right of the building, but

disconnected, a very high minaret was added, the features of which betray

its date. It disappeared after the First World War but is known from old pho-

tographs. The han opposite the zaviye-mosque disappeared in the nineteenth

century. The hamam, however, is preserved and was excellently restored in the

1980s. The remarkable disrobing room is covered by a dome on an exceedingly

high tambour, a characteristic feature of the early Ottoman dome and very

probably related to the stilted domes of late Byzantine architecture. The plan

of both buildings was evidently imported from Anatolia but the workman-

ship of the walls, with rosettes in brick and other ceramo-plastic elements, is

related to the local architecture.

Related to the buildings in G

ulcine and

Ihtiman is the long misunderstood mosque of Yıldırım Bayezid in the western-

most suburb of Edirne. This building is, like that in G

ulcine and the older

zaviyes in Tokat and elsewhere, not oriented towards Mecca. It was therefore

thought to have been a Byzantine church of unknown type. It has also been

interpreted as not being from Yıldırım Bayezid but from Murad I, thus decades

older, on the grounds that Edirne had no sultanic mosque. That at least two

large ex-churches served as mosques, as is related by Hibri and Evliya C¸ elebi,

74 For this buildingsee SemaviEyice, ‘Sofya Yakınında

Ihtiman’da Gazi Mihalo

glu Mahmud

Bey

Imaret-Camii’, Kubbealtı Akademi Mecmuası 4, 2 (1975), 49–61. For the town and its

district see M. Kiel, ‘

Ihtiman’ in TDV

IA, vol. XVIII (Istanbul, 1999), vol. XXI (2000),

pp. 571–2.

167

machiel kiel

was overlooked. The building is traditionally called Yıldırım Cami, or, more

revealing, Eski

Imaret, to distinguish it from the Orta

Imaret of Mihalo

glu

Mehmed Bey and the great Yeni

Imaret of Bayezid II, likewise on the banks

of the Tunca. It is also mentioned in some ﬁfteenth-century account books

of great sultanic foundations, showing a staff including two cooks, one baker,

one dishwasher, a cellar-master and a rice pounder for the food distributed

in the building.

Moreover, its original inscription from 1399–1400, preserved

broken into several pieces and kept in the Edirne Archaeological Museum

(Inv. No. 1537), calls the building an imaret, not a mosque.

The plan shows a

central dome on the intersection of four barrel-vaulted eyvans (recesses open

on one side). Two of the remaining corners are ﬁlled with detached rooms,

furnished with chimneys and opening to the main building through doors.

They were evidently the guest rooms. In front of the building, on the north-

ern side, there once ran a porch of ﬁve units of which only the columns are still

standing. Outside the building are the ruins of the kitchen building, with one

huge chimney still standing. Only at a much later date was this original zaviye,

called imaret in the inscription, transformed into a mosque proper, with the

addition of a minaret and a mihrab obliquely placed in a corner of the southern

eyvan in order to show the correct orientation to Mecca.

Another building that has long puzzled art historians is the curious mosque

of C¸ elebi Sultan Mehmed (1413–21) in the small town of Didymoteichon (Dime-

toka). This town was conquered before Edirne and had long received special

treatment from the Ottoman sultans. Mehmed the Conqueror and Bayezid II

were born in the palace of Dimetoka and in the ﬁfteenth century the imperial

treasure was kept there. The great mosque which still overlooks the town

was begun by Yıldırım Bayezid before 1402 but completed much later by his

youngest son, Mehmed I. The Ottoman registers from the late ﬁfteenth cen-

tury onward call it the mosque of Sultan Yıldırım Han, as do Evliya C¸ elebi

and the Ottoman accounts of the

Imaret-i Bayezidiye in Edirne.

The Dime-

toka mosque was conceived as an enlarged version of the S¸ehadet Cami in

the castle of Bursa, a work of Murad I, with two large domes resting on two

massive piers over the central ‘nave’ and two mighty cradle vaults on each of

Omer L

utﬁ Barkan, ‘Edirne ve Civarındaki Bazi

Imaret Tesislerinin Yıllık Muhasebe

Bilanc¸oları’, Belgeler 1, 2 (1964), 235–377 (for the Bayezidiye

Imaret, pp. 291–3).

76 For this important inscription see the scholarly edition by F. Th. Dijkema, The Historical

Monumental Inscriptions in Edirne (Leiden, 1977), pp. 13–15 and plate 1.

77 Evliya C¸ elebi, Seyahat-namesi, vol. VIII (Istanbul, 1928), p. 75. See Barkan, ‘Bazı

Imaret

Tesislerinin’, p. 292.

168

Ottoman expansion into the Balkans

the remaining lateral sections.

When the walls had reached their full height

Bayezid was defeated by Timur at the battle of Ankara (1402), and the building

remained unﬁnished. Almost two decades later Sultan Mehmed I ﬁnished the

work in a different and much cheaper manner than originally intended. As

he had already completed the huge Eski Cami of Edirne, left unﬁnished by

his brothers S

uleyman and Musa, and was still working on the very expensive

Ye s¸il Cami in Bursa, he evidently had to economise. Instead of stone the entire

vaulting was made of wood and covered by a huge, lead-covered, pyramid

roof. The woodwork, already described by Evliya in 1669, is not the product

of later repairs but dates from 1420–2, as was shown by dendrochronological

research by Striker and Kuniholm. In the long and elaborate inscriptions over

the entrances only Mehmed is mentioned as founder, as he is in the inscrip-

tions of the Edirne Eski Cami. The provisional character of Mehmed’s work

in Dimetoka can also be seen in the fact that the mosque never had a vakıf of

its own. Its staff were paid from the poll-tax of the Christians of the kaza of

Dimetoka.

In spite of this, the Dimetoka mosque remains one of the largest

architectural enterprises of the Ottomans in the Balkans outside the capitals

Bursa and Edirne. It is also the ﬁrst building in the Balkans with an inscription

where the architect mentions himself. This is the well-known Hacı

Ivaz Pas¸a

of Tokat, the architect of the famous Yes¸il Cami in Bursa. It is thus an unde-

niable document of the way Ottoman architecture was introduced into the

Balkans from Anatolia. The building now has no function but is maintained

by the Greek Service for Byzantine and Post-Byzantine Antiquities, Kavalla.

The little Thracian town has more early Ottoman buildings, mostly very lit-

tle known but still standing. One of them is the so-called Fısıltı Hamamı, or

Whispering Bath, a foundation of Uruc¸Pas¸a, one of the sons of Timurtas¸,

the second beylerbey of Rumeli.

Hibri Efendi in the 1630s, and Katip C¸ elebi

and Evliya C¸ elebi some decades later, mention this hamam. Hibri Efendi, who

himself had worked at the medrese of Uruc¸Pas¸a, gives the date of construction

of this early example of an Ottoman hamam in the Balkans as 801/Septem-

ber 1398–September 1399. Like other early Ottoman hamams it has no great

domed disrobing room, and never had one, as can be seen on the ﬁnely worked

78 See Ayverdi’s reconstruction in Ekrem Hakkı Ayverdi, Osmanlı Mim

ısinde C¸ elebi ve II.

Sultan MuradDevri806–85 5 (1403–1451), vol. II (Istanbul, 1972)pp.136–50. The inscriptions

of this building are also given here in detail.

79 See references in various Vazifehoran Defters in the BOA (i.a. mad 5625 from 1047–1637,

p. 4; also mad 6177 from 1046).

80 For the town in the Ottoman period see M. Kiel, ‘Dimetoka’ in TDV

IA, vol. IX (1994),

pp. 305–8.

81 Mehmed S

ureyya, Sicill-i.‘Osm

ı, vol. I (Istanbul, ah 1308), pp. 422–3.

169